United States
                   Environmental Protection
                   Agency
Water Engineering
Research Laboratory
Cincinnati OH 45268
                   Research and Development
EPA/600/S2-85/094 Sept. 1985
&ER&         Project Summary
                   Pilot Study for  Removal of
                   Arsenic  from  Drinking
                   Water  at  the  Fallon, Nevada,
                   Naval Air Station

                   Frederick Rubel, Jr., and Steven W. Hathaway
                     This report presents the pilot-plant
                   test results of two treatment methods
                   for removing arsenic from drinking
                   water—activated alumina and ion ex-
                   change. A mobile trailer was placed at a
                   site near the water distribution system
                   on the Fallen, Nevada, Naval Air Station
                   (NAS) grounds, where the arsenic con-
                   centration was measured at 0.080 to
                   0.116 mg/L. This level exceeded the U.S.
                   Environmental Protection Agency
                   (EPA) maximum contaminant level
                   (MCL) of 0.050 mg/L.
                     The trailer was equipped with three
                   PVC testing columns and an analytical
                   laboratory for the  pilot project. The
                   NAS drinking water was used for evalu-
                   ating the efficacy of treatment under
                   several different conditions. The acti-
                   vated alumina and  ion exchange sys-
                   tems were operated through three
                   loading and regeneration cycles each.
                   The major water quality factors affect-
                   ing the removal of arsenic by these
                   methods were pH of feed water, arsenic
                   concentration, suKate  concentration,
                   and alkalinity. The  major operational
                   factors affecting removal were flow
                   rate, down time, and media clogging.
                   The report also estimates the capital
                   and operating costs for arsenic removal
                   using the activated alumina method at
                   optimum pH (5.5) for each of the three
                   small community systems currently
                   using water from the same aquifer. The
                   pilot study estimated NAS capital costs
                   of $558,000 for a 1-mgd plant and oper-
                   ating costs of 220/1,000 gal of treated
                   water.
  The project report also addressed
treatment and handling of the waste-
water generated by the arsenic removal
process, an issue often omitted from
treatabilrty studies. Several containers
of the regeneration waste were used
for a special study to characterize, de-
water, and render the waste nontoxic
for disposal in a sanitary landfill.
  This project summary was developed
by EPA's Water Engineering Research
Laboratory, Cincinnati, OH, to an-
nounce key findings of the  research
project that is  fully documented in a
separate report of the same title (see
Project Report ordering information
at back).

Introduction
  Many small communities are faced
with contamination of their ground-
water by inorganic chemicals. Arsenic
contamination can result from the
leaching of manmade toxic compounds
into the groundwater, or it can be
caused by the  natural dissolution of
minerals from subterranean strata. Re-
gardless of the source of contamina-
tion, most small communities that have
drinking water with arsenic concentra-
tions higher than the U.S. Environmen-
tal Protection Agency (EPA) maximum
contaminant level (MCL) of 0.05 mg/L
do not have existing treatment facilities
that can be modified to reduce the ar-
senic to acceptable levels.
  The arsenic  concentration  in the
Fallen, Nevada, Naval Air Station (NAS)
water supply was measured at 0.080 to

-------
0.016 mg/L, which exceeded the MCL of
0.050 mg/L established by EPA and the
Nevada State Health Division of the Bu-
reau of Consumer Health Protection
Services. A pilot plant study was, there-
fore, conducted to evaluate arsenic re-
moval from drinking water at the Fallon,
Nevada, Naval Air Station (NAS)  using
activated alumina  and ion exchange
technology.  The primary objective of
the test project was to develop informa-
tion that would lead to the design and
economical installation of cost effective
treatment  systems  that could  provide
potable water  in compliance with the
MCL for arsenic.
  Although the NAS, the City of Fallon,
and the Fallon  Indian Reservation were
the primary  recipients  of this informa-
tion, the research data were developed
for use by other small  communities in
determining  cost effective treatment al-
ternatives for  removing arsenic from
drinking water.
  The work plan required a series of ar-
senic removal treatment tests. The stud-
ies included data acquisition and evalu-
ation  of three separate treatment
modes using two different media. The
three treatment processes were as fol-
lows:
  A. granular  activated  alumina with
     pH adjustment (pH 5.5),
  B. granular  activated alumina with-
     out pH  adjustment, and
  C. strong-base anion exchange resin
     without pH adjustment.
  Each treatment process was evalu-
ated through three  complete treatment
cycles, with  a  cycle consisting of one
treatment run and one regeneration.
Additional tests were  carried out to
study the  effects of activated  alumina
treatment at  pH 6.0 and using upflow at
pH 5.5.
  The  treatment  media used  were
(1) Alcoa F-1*  granular  activated alu-
mina (28  to 48 mesh), and (2) Dow
Chemical Company Dowex SBR strong-
base anion resin (20 to 50 mesh).
Test Apparatus
  The pilot plant apparatus consisted of
three separate systems—one for each
treatment  mode and piped in parallel
(Figure 1). Each pilot treatment system
was designed to treat the well  water at
a maximum  flow rate  of 1-1/2 gpm at
50 psi maximum working pressure. The
chlorinated well water  was pumped to
the test site by NAS through a 1-in.,
 "Mention of trade names or commercial products
 does not constitute endorsement or recommenda-
 tion for use.
PVC, schedule 80 pipe  at an approxi-
mate pressure of 75 psig (the maximum
was 150  psig,  and the  minimum was
50 psig). The water flowed through a 5-
u,, powdered, activated carbon cartridge
filter for removal of chlorine, through a
pressure control valve, into a 1-in. man-
ifold. The manifold branched into three
separate systems,  each with one 10-in.-
diameter by 60-in.-high treatment
column containing 1.0 ft3 of treatment
media.
  The treatment columns had remov-
able heads on top  and bottom. The
underdrain section occupied the bottom
6 in. of each column, and the treatment
media filled 22 in. The remaining height
was available for media expansion and
freeboard. The test apparatus  was
piped  for manual operation, with  each
treatment system independent of the
other two.
  The entire treatment system  was
mounted  in an insulated laboratory
trailer 30 ft long by 8 ft wide by 8 ft high.
In addition to the pilot plant equipment,
the trailer housed a chemical laboratory
that was equipped to perform all analy-
ses required  during the  duration of the
test program.

Pilot  Test Program
  The 1 ft3 of treatment media  was
placed in  each treatment vessel. The
volume was determined by weight,
using the manufacturer's  published
data for material density (Alcoa F-1  52
Ib/ft3;  Dowex SBR 44 Ib/ft3). Bed vol-
umes  were also measured. The media
were carefully backwashed for removal
of fines.
  The first cycle  for each  treatment
method was started at approximately
the same time. Initially, each mode
treated the water at the same flow rate,
1-1/2 gpm. This flow rate was estab-
lished as an  optimum during previous
pilot test programs. Increasing the flow
rate may lower the removal  efficiency;
decreasing the flow rate may or may not
increase efficiency, but  it does extend
the test duration. This effect was moni-
tored  closely during the testing. For
Column A, the treatment runs were ex-
tremely long and permitted iron de-
posits to build up in the media. This
buildup resulted in loss of treatment ef-
ficiency. Thus  after  40  percent of the
first treatment run, the flow rate was
permanently decreased to 1  gpm. This
change alleviated the iron problem and
improved the treatment efficiency. The
runs for  Columns B and C  were very
short, and thus no flow rate adjust-
ments were considered for those
modes. After the initial startup, no at-
tempt was made to synchronize the be-
ginning or end of cycles of the different
treatment modes. Pilot test operation
for all practical  purposes was continu-
ous; one cycle  immediately followed
the completion of the previous cycle for
each treatment mode.
  Raw and treated water samples for
each treatment mode were collected at
least once per day. During each regen-
eration, backwash  and regeneration
wastewater grab samples  were col-
lected at  5- and 10-min intervals. Grab
samples were also collected each time
the wastewater  pH dropped 1 unit dur-
ing regenerations in which there was a
neutralization step. Composite samples
were taken during each step of regener-
ation. Analyses performed for  raw,
treated,  and regenerated water are
listed in Table 1. A complete raw water
analysis appears in Table 2. Since all
treatment  methods were evaluated
using chlorinated feed water, the ar-
senic was assumed to  be in the +5 va-
lence form. This assumption is based on
known arsenic water chemistry and  is
supported by a  recent investigation.
  Each treatment method was operated
independently  during testing. A de-
scription of the operation of each
method follows.

Method A. Activated Alumina
with pH Adjustment

  The raw water was treated in a down-
flow, packed-bed  configuration. The
flow rate was controlled initially at 1-1/2
gpm and then reduced to 1  gpm at 40
percent through the first run. Before en-
tering the  treatment vessel, the raw
water was  adjusted to  pH 5.5 using  a
dilute solution of sulfuric acid (H2SO4).
  Each treatment run extended until the
arsenic level in the treated water was at
or very near the arsenic level in the raw
water. The treatment bed was  then
backwashed with raw water, drained,
and regenerated in the  upflow direction
with 4 to 4.5 percent NaOH solution. The
bed was then  regenerated again in  a
downflow direction. Upon completion
of the downflow regeneration, the bed
was  flushed with  raw  water at a flow
rate of 1-1/2 gpm for 15 min.
  At this point,  the  raw water pH was
adjusted  to 2.5 and the neutralization
phase of the regeneration began. The
effluent pH was higher than 12 at the
beginning of this phase, and it dropped
slowly over an extended period. When

-------
 IXJ

 EH

 -C
  Legend

- Interface with Facility Piping

- Pressure Indicator

- Check Valve

- Ball Valve
- Hose Connection
                        Dilute HaSO
                         Day Tank
                       0-160  0-60
                        psig  psig

                        O OP'
                   -vat-
                       Pressure
              Carbon   Reducing
             Cartridge     Va/ve
               Filter
            f Chlorine Kill)
   4% NaOH   10%NaCR
    Batch      Batch
     Tank      Tank
1
o
0-60 psig
Turbine (~) p/
Meter V
m _ ?
r'\V2" f '-V'ifA
-T Rate of -*t- 5
f/oiv
Controller
Column 'A'
Granular
A ctivated A lumina
with pH A djustment
A cid Feed for Neutralization
.After Regeneration Q_^Q
I Turbine (~) PI
f Merer j =
M 8 U-V* ~l _ T S
^"/?ate/7ute /VaOW
Day Tank
^Sample
0-60 psig
Qpi 1
4 T Treated Effluent
Sample |^
3
-. Sample
0-60 psig
Op,
I Treated Effluent
4V" * l'/2
" Sample *YDr
3
-. Sample
O-60 psig
On
| Treated Effluent
3 A/"|^" 1 1 ^
                                                                                Sample  \Drain
                                                                                                   J'/2"
                                                                                                 Drain to Waste
Figure 1.    Flow diagram for arsenic removal pilot plant.
the effluent pH reached 9.0, the influent-
adjusted pH was raised from 2.5 to 4.0,
at which point the next treatment run
began. When the effluent pH dropped to
6.5, the influent-adjusted pH was raised
from  4.0 to 5.5,  where  it remained
throughout the treatment run. The efflu-
ent pH was then  adjusted to 7.5  with
dilute caustic. Because of the cost of
chemicals and because adding sodium
is  not desirable, aeration  was used to
raise the pH of the effluent for the re-
mainder of  the study.
                                  Method B. Activated Alumina
                                  without pH Adjustment

                                    This method was identical to Method
                                  A except that pH was not adjusted dur-
                                  ing the treatment run. The only pH
                                  change occurred during the neutraliza-
                                  tion step after regeneration.
                                    Backwash and regeneration proce-
                                  dures were also the same as for Method
                                  A except that the acid feed adjusting the
                                  pH to 2.5 was terminated when the ef-
                                  fluent pH dropped to 9.0. The next treat-
ment run began then. The pH was not
adjusted further until the neutralization
phase of the next regeneration.

Method C. Strong Base Anion
Exchange Resin without pH
Adjustment

  The raw water was treated in a
downflow, packed-bed configuration at
1-1/2 gpm. No pH adjustment was
made during the treatment  or after
regeneration. Regeneration with 10%

-------
NaCI was used with the anion resin
followed by treated water rinse.

Results of Pilot Tests

Treatment Method A,
Activated Alumina with pH
Adjustment
  The three test cycles of Method A ex-
periments were labeled A1, A2, and A3,
respectively. Each cycle ran for more
than 3 months before media exhaus-
tion.
  Figure 2 presents loading curves for
the first run of each method tested (A, B,
and C). For cycle A1, arsenic was not
detectable in the effluent up to a treat-
ment volume of about 8,500 bed vol-
umes (BV) (63,000 gal). The effluent ar-
senic concentration reached the MCL of
(0.05 mg/L) at  15,536 BV (116,210 gal).
The media continued to remove arsenic
up to 24,500 BV (183,600 gal). At this
point,  effluent arsenic concentration
reached 0.085  mg/L. The  test run was
terminated  because the activated alu-
mina capacity was nearly  exhausted.
  The  arsenic removal capacity has
been calculated for all test runs  (Table
3). In Test Run A1, the alumina removed
325 grains/ft3 (747 g/M3) of arsenic be-
fore any breakthrough occurred. The
test run was continued until the effluent
arsenic concentration reached the influ-
ent concentration. Total capacity calcu-
lated at this point was 655 grains/ft3
(1507 g/M3). Table  4 displays the test
data in terms of volume and run time.
  The  average amount of acid  (66°B'
sulfuric acid) used to adjust raw water
to pH 5.5 during Method A testing was
0.130 gal/1000 gal of treated water.
  During Test Run A1, the potential ex-
isted for blending  50 percent treated
water with  50 percent raw water and
still complying with the arsenic MCL.
Since the raw water arsenic level in later
treatment  runs consistently  exceeded
0.110  mg/L, a conservative design
would  provide for blending 75 percent
treated water with 25 percent raw water.
Thus a water with 0.03 mg/L arsenic will
provide more flexibility in cases  where
treatment  conditions may change.
Table 5 analyzes the 75/25  blend for Run
A3 at 3,979 BV (29,760 gal).

Treatment Method B. Activated
Alumina with No pH
Adjustment
  Column B contained 1 ft3 of F-1 Alcoa
activated alumina, as in Column A, and
operated at 1.5 gpm. The  raw water pH
Table 1.    Analytical Tests for Water Samples

           Item
 Daily Tests
Weekly Tests
 Treatment Methods A & B
 (activated alumina):
Raw and Treated Water






Regeneration Wastes
(samples taken at each re-
generation)



Treatment Method C
(anion exchange)





pH
Arsenic
Fluoride
Aluminum
Silica
Alkalinity
TDS
Arsenic
Aluminum
pH
Fluoride
Chloride
Sodium
pH
Arsenic
Alkalinity
TDS
Sulfate
Chloride
Fluoride
Total hardness
Carbonate
Bicarbonate
Sodium
Sulfate
Color

Calcium
Magnesium
Sulfate
TDS
Alkalinity
Total hardness







of 9.0 was not adjusted before treat-
ment in the three runs (B1, B2, B3). The
loading curve for Run B1  is compared
with those for A1 and C1 in Figure 2.
  Test Runs B1, B2, and B3 were  all
short cycles because of low removal ef-
ficiency. At pH 9.0 to 9.1, alumina is un-
favorable for arsenic removal but ideal
for silica removal. As expected, the  ar-
senic removal  performance was very
poor. Arsenic  removal was complete
after only 3 days for Run B1, but Runs
B2 and B3 stopped removing arsenic af-
ter only 1 day. The raw water arsenic
level was 0.090 mg/L during the test pe-
riod. The treated water arsenic level was
as low as 0.008 mg/L. Because only one
Table 2.    Fallon NAS Raw Water Analysis
water sample was collected per day,
there are very few data points. How-
ever, extrapolation of available data in-
dicates that during the run, 800 BV (5940
gal) were treated  and 19 grains of ar-
senic were removed. Tables 3  and 4
summarize the results of the loading
curve. The treatment run was continued
until silica removal terminated.
  For a  short period after arsenic re-
moval ceased, a small amount of ar-
senic was desorbed from the bed. This
result indicates that at the higher treat-
ment pH, the alumina prefers silica to
arsenic.  Some measurable fluoride re-
moval was also detected during the first
day of the run.
Analyte
Total alkalinity (as CaCo3)
Hardness (as CaCo3)
Calcium
Magnesium
Sodium
Aluminum
Chloride
Sulfate
Fluoride
Arsenic
TDS*
pH (units)
SiO2
Mean
246
5
1
1
240
0
101
96
0.788
0.103
535
9.1
28
Range
228-263
4-6
0-2
1
223-266
0
87-120
88-120
0.70-1.10
0.080-0.116
495-560
9.0-9. 1
28-34
"Values in mg/L except as noted.
f Total dissolved solids.

-------
 Treatment Method C. Ion
 Exchange
  The results of the ion exchange tests
 of Method C were similar to those of
 alumina with  no pH adjustment in
 Method B. During the initial work with
 the strong-base resin, it became imme-
 diately obvious that sulfate was pre-
 ferred to arsenic. Since  the raw water
 contained almost 100 mg/L sulfate and
 less than 0.1 mg/L arsenic (ratio 1000:1),
 the resin was expected to have nearly
 zero capacity for arsenic. And because
 sulfate is preferred, sulfate will displace
 the arsenic after the arsenic  break-
 through point.  Alkalinity was also re-
 moved during the early part of the  run
 as shown by dropping pH; but later it
 was also eluted  from the bed by the sul-
 fate. Figure 2 shows the loading curve
 for Run C1 compared with those for B1
 and A1. Only 300 to 500 BV (2244 to
 3740  gal)  could be  treated before  the
 resin was exhausted for arsenic re-
 moval. Tables 3 and 4 show the break-
 through data points given for Runs C1,
 C2, and C3.

 Testing  under Non-Optimum
 Conditions
  Because of the short test runs of alu-
 mina at pH 9.0  (Runs B1, B2, and B3),
 two tests  were  conducted to compare
 the performance of an  activated alu-
 mina run at pH  6.0 with the test  runs at
 pH 5.5 (Runs A1, A2, and A3). Thus,
 Column B was loaded with a fresh bed
 of virgin activated alumina identical to
 that used in Column A. If the tests were
 successful, the benefits of operating at
 pH 6.0 instead of pH 5.5 would include
 the following:
  1. Use of less acid for pH adjustment,
  2. Higher pH  of blended water, and
  3. Lower treated water sulfate level.
  The initial flow rate was 1.4 gpm in a
 downflow configuration, similar to that
 of Method A. After approximately 40
 percent of the  run,  the flow rate was
 reduced to 1.0 gpm.
  Run B4 was terminated after treating
 17,400 BV (130,000 gal) for a 29 percent
 decrease in capacity from Run A1 at pH
 5.5 (24,500 BV, or 183,260 gal). Figure 3
 compares the results of Run B4 with
 Run A3. No arsenic was detected in the
effluent until 7000 BV (52,360 gal) were
treated. A  rather steady increase was
then detected up to about 12,000  BV
 (89,760 gal). At 12,000, 15,000,  17,000
and 19,000 BV (89,760,112,220,127,160,
 142,120 gal), the alumina bed seemed to
 recover some capacity. This recovered
Table 3.    Summary of Column Capacity
                          Arsenic Removal Capacity

                                       Grains/ft3 (g/M3)
Test Run ID
A1 (Virgin)
A2
A3
B1 (Virgin)
B2
B3
C1 (Virgin)
C2
C3
B4 (Virgin)
B5
At First
Arsenic Detection
325 (747)
42 (95)
246 (566)
N.C."
2.6 (5.8)
2.6 (5.8)
N.C."
5.7 (13.0)
9.9 (22.4)
241 (554)
Df
At
Arsenic MCL
540 (1242)
490 (1107)
646 (I486)
18.3(42.1)
5.0(11.3)
3.2 (7.2)
N.C."
10.1 (22.8)
13.8 (31.2)
433 (979)
294 (676)
At
Media Exhaustion
655 (1507)
611 (1405)
730 (1679)
19.3 (44)
5.2(11.0)
3.7(8.5)
8.2 (18.5)
10.2 (23.)
13.9 (31.4)
536 (1233)
488 (1122)
"Samples not collected at low bed volume.
*D—All samples contained detectable arsenic.

Table 4.    Summary of Treatment Volumes and Run Time

                       Volume Treated, galfbed volumes)
Test Run ID
A1
A2
A3
B1
B2
83
C1
C2
C3
B4
B5
At
First Arsenic
Detection
63,855 (8542)
8,270(1105)
39,890 (5332)
756 (101)
756 (101)
1,590 (212)
1,350 (180)
2,160(288)
35,550 (4752)
At
Arsenic MCL
116,210(15,536)
96,400 (12,888)
111,700(14,933)
5,940 (794)
1,260 (168)
2,160(288)
100,165(13,391)
57,560(7.695)
At
Media
Exhaustion
183,600 (24,545)
152,900 (20,447)
158,870 (21,239)
7,990 (794)
1,764 (235)
1,008 (134)
3,695 (493)
2,430 (324)
2,970 (397)
130,080 (17,390)
102,020 (13,639)
Run Time
(days)
111
105
110
4
1
1
1
1
1
94
67
Table 5.   Analysis of Water from Run A3—Activated Alumina with Raw Water pH Ad-
         justed to 5.5



Characteristic
P Alkalinity (as CaCO3)
M Alkalinity (as CaCO3)
Hardness (as CaCO3)
Calcium
Magnesium
Sodium
Aluminum
Chloride
Sulfate
Fluoride
Arsenic
Silica
TDS
pH (units)


Raw Water
(mg/L)'
50
246
5
0
1
240
0
101
96
0.80
0.110
28
535
9.1


Treated Water
(mg/L)"
0
30
5
0
1
240
.02
101
290
0.55
NDf
28
630
5.5

Aerated
Treated Water
(mg/L)"
0
30
5
0
1
240
.02
101
290
0.55
ND
28
630
7.5
Blend of 75%
Aerated Treated
Water with 25%
Raw Water
7
74
5
0
1
240
.02
101
230
0.65
0.028
28
600
8.6
"Except for pH.
tNot detectable.

-------
    Q.12-\
    0.10-
t
I  o.os-
Ui
    0.06-
    0.04-
    0.02-
    0.00-
                    5000
                                 ~r
                                10000        15000
                                   Bed Volumes
                   20000
25000
Figure 2.    Loading curies for Run 1. methods (A.B.C).
capacity  only occurred when the
column was shut off for a short period
of time. The data in Tables 3 and 4 show
the capacity of the alumina at various
points  in the run. Before arsenic detec-
tion in the effluent, the alumina re-
moved 241 grains/ft3. At exhaustion, ar-
senic removal capacity for B4 was 536
grains/ft3 versus 655 grains/ft3 for Run
A1—an 18 percent reduction.
  Before regeneration, the adjusted raw
water pH was lowered to 5.5 to deter-
mine whether any additional arsenic
could be removed by the alumina. The
results showed that  an  additional 26
grains  of  arsenic was removed from
2,580 BV (19,300 gal) at the lower pH.
  Acid  consumption for the reduction of
raw water pH to 6.0 was 0.107 gal 66°B'
sulfuric acid versus 0.130 gal for the pH
adjustment to 5.5 (17 percent reduc-
tion). Though this mode could reduce
acid consumption slightly and reduce
the sulfate concentration in the treated
water,  these  benefits were  offset by
more frequent treatment-bed regenera-
tions. The latter liability more than off-
sets the benefits  by increasing operat-
ing  labor costs,  generating more
wastewater,  and  producing higher
chemical  costs for regeneration. Thus,
although  operating with  treatment pH
at 6.0 is technically feasible, it is offset
by  economic disadvantage and in-
creased wastewater disposal volume.
  Run  B5 was carried out in the upflow
direction at 1.0 gpm, with raw water pH
adjusted to 5.5. This  experimental run
provided a comparison of upflow and
downflow treatment of a regenerated
bed at pH 5.5, the same pH used in Runs
A2 and A3. A successful result could re-
duce the wastewater volume and elimi-
nate the requirement for backwashing
the bed. However, treatment capability
compared unfavorably with the down-
flow mode. The treated water arsenic
level did not reach the undetectable lev-
els  common to all downflow arsenic
treatment  runs; only a small amount of
    0.12 H
    0.10-
    0.08
treated water—2,400 BV (18,000 gal)—
had levels lower  than 0.010  mg/L
(Figure 3). The treated water arsenic lev-
els  initially rose above the MCL after
treatment of only 8,000 BV (59,200 gal)
compared with 13,000 BV (97,800 gal)
for  Run A2 and 14,800 BV (113,100 gal)
for Run A3. The results suggest that the
upflow operation is far more vulnerable
to channeling. After various methods
were attempted for  restoring arsenic re-
moval capacity to a spent bed, the run
was terminated. The total arsenic re-
moved was 488 grains after 21,308 BV
(157,080 gal) were treated.

Regeneration  of Treatment
Media
  When the activated alumina columns
became saturated  with arsenic, they
were regenerated with a 4 to  5 percent
NaOH solution. The procedure for re-
generation included upflow treatment,
then downflow treatment, raw  water
rinse, and finally  neutralization with
H2S04. Method A showed some reten-
tion of arsenic on the bed. A mass bal-
ance calculation gave an overall recov-
ery of 80 percent for the three Method A
regenerations. Approximately  1,996
grains of arsenic were loaded  on
Column A, and three separate regenera-
tions recovered 1602 grains. However,
the results of  analysis of the concen-
trated waste stream are questionable
because of high TDS in the wastewater.
No  significant lowering of capacity was
noted from Run A1  to A3 (Table 3).
 I
    0.06
    0.04-
    0.02-
    0.00
            MCL
                                        ___
-------
  Column  B was  regenerated in  the
 same manner as Column A. Total load-
 ing of about 23 grains of arsenic and
 recovery of 19 grains yielded a recovery
 of 82 percent. The ion exchange resin
 (Column C) was regenerated  with 10
 percent NaCI solution in the downflow
 mode only. Total  loading was 32.3
 grains, and 17.7 grains were recovered,
 producing a recovery of only 55 percent.

 Handling of Waste Regenerant
  The regeneration of spent arsenic sat-
 urated media produced a waste product
 high in dissolved solids, aluminum, and
 soluble arsenic.  Analysis of the waste
 regeneration product from activated
 alumina treatment revealed an arsenic
 concentration range of 23 to 41 mg/L
 and a pH of 12.  EPA classifies a waste
 stream with 5 mg/L or more of soluble
 arsenic as  a toxic waste that must be
 disposed of in an approved hazardous
 and toxic waste landfill.
  A limited investigation was con-
 ducted to treat the arsenic-laden waste
 and render it nontoxic. This procedure
 involved precipitation of the arsenic
 with the aluminum already present in
 the wastewater by adjusting the pH to 5
 to 6.5, at which point the aluminum hy-
 droxide precipitates out of solution. The
 solids were separated from the super-
 natant by mechanical dewatering. Anal-
 ysis of dry  solids by the EPA extraction
 procedure  (Federal Register, Vol.  45,
 No. 98, Monday May 19,1980, Appendix
 II—EP Toxicity Test Procedure) showed
 that the dry cake from the centrifuge
 contained  1627  mg arsenic/kg solids.
 The liquid  extract from the dry solids
 contained  only 0.036 mg/L arsenic,
 which is well within the 5-mg/L limit.

 Cost Estimates
  Cost estimates were calculated for the
 three small communities based on the
 results  of  Method A treatment (acti-
 vated alumina with raw water pH ad-
 justed to 5.5). The design flows for the
 three systems are 700 gpm for the NAS,
 2400 gpm for the City of Fallon, and 100
 gpm for the Fallon Indian Reservation.
 Capital costs  were estimated at
 $558,000, $1,343,000, and $179,000, re-
 spectively.  Operation and maintenance
 costs  per 1000 gal  treated were 22.10,
 31.80, and 29.80, respectively. The high
costs for the City of Fallon were the re-
 sult of the need to modify the distribu-
tion system and for high pumping costs
associated with fire prevention require-
ments.
Conclusions
  Removal of the pentavalent arsenic
from the Fallon NAS drinking water was
best achieved by the activated alumina
system treating the raw water adjusted
to pH 5.5 with sulfuric acid. Fluoride re-
moval occurs along with arsenic  re-
moval. Blending raw and treated water
will allow some fluoride to be present in
the product water. The  waste stream
from  regeneration of spent activated
alumina could be classified  as  toxic
since soluble arsenic is greater than 5
mg/L. However, the arsenic can be pre-
cipitated from the waste with high in-
strinsic aluminum content by lowering
the pH to 5 to 6.5. This process results in
a supernatant that is very low (<0.1
mg/L) in soluble arsenic and a solids
portion in which the arsenic  is non-
leachable according to the EPA extrac-
tion procedure toxicity test.  Use  of
strong-base anion exchange resin was
inefficient for arsenic (+5) removal be-
cause of the competition of the high sul-
fate concentration.
  The full report was submitted in fulfill-
ment of IAG AD-17-F-3-481-0 by U.S.
Department of the Navy under the spon-
sorship of the U.S. Environmental Pro-
tection Agency.
   Frederick Rubel, Jr. is with Rubel and Hager, Inc.. Tucson, AZ8S711; and Steven
     W. Hathaway  (also the EPA  Project Officer,  see below) is with  Water
     Engineering Research Laboratory, Cincinnati, OH 45268.
   The complete report.
     Water at the Pallor
 entitled "Pilot Study for Removal of Arsenic from Drinking
, Nevada, Naval Air Station," (Order No. PB 85-243 178/AS;
     Cost: $11.95, subject to change) will be available only from:
          National Technical Information Service
          5285 Port Royal Road
          Springfield. VA 22161
          Telephone: 703-487-4650
   The EPA Project Officer can be contacted at:
          Water Engineering Research Laboratory
          U.S. Environmental Protection Agency
          Cincinnati, OH 45268
                                                                               U. S. GOVERNMENT PRINTING OFFICE:1985/559-l 11/20694

-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
    BULK RATE
POSTAGE & FEES PAII
       EPA
  PERMIT No. G-35
Official Business
Penalty for Private Use $300

EPA/600/S2-85/094
                0169064   WERL

                US  EPA REGION  V
                LIBRARY
                230 S  DSAR80RN  ST.
                CHICAGO              IL
             60604

-------